Research And Application Of Noise Processing In 3D Laser Measurement System

With the rapid development of modeRN science and technology,3D measuring technology is also constantly improved from the traditional 3D measuring machine to the optical measuring instrument which having the advantages of non-contact,fast,easy to automate and so on.It has wide application in the fields of industrial inspection,biomedicine,product quality control and reverse engineering.The processing of large discrete data is becoming more and more important,and the smoothing of point cloud based on discrete data has become a hot research topic in recent years.This thesis analyzes the error and error sources in 3D measured data,including the random error of the machine itself and the machine thermal noise caused by vibration,error caused by the laser line width,machine assembly precision and surface roughness of the object,color,texture and other factors.Then it summarizes the research of point cloud data processing at home and abroad,with emphasis on curvature calculation and smoothing algorithm based on curvature.In this thesis,based on the differential geometry,some basic concepts of curves and surfaces are introduced,and then the formulas of Gauss curvature and mean curvature are given.In this thesis,we introduce the concept of smoothing noise,the criterion of fairing,the principle of fairing and some methods of fairing at home and abroad.The properties of mean curvature flow are also introduced emphatically.This thesis introduces the calculation method of curvature based on Mark Mayer algorithm.By improving the calculation of the Voronoi area and improving the calculation method of Mark Mayer curvature,we design improved Mark Mayer curvature algorithm,then the algorithm is applied to improve the mean curvature flow data smoothing algorithm.Compared with other smoothing algorithms,the improved algorithm can better deal with the surface detail features,objective to achieve fairness at the same time,keeping the surface geometric features invariant.According to this algorithm,my laboratory develops a 3D scanning software,mainly realizes the data acquisition,data display,data processing,data interface module.It can extract the image precisely and remove noise in the scanned data effectively.It has been recognized by relevant enterprises.It is beneficial to the popularization of 3D scanning technology and it makes contribution to the development of science and technology.